Stephen Frost <sfrost@snowman.net> writes:
> * Tom Lane (tgl@sss.pgh.pa.us) wrote:
>> One thing that *would* be worth doing is merging upstream's recent
>> updates; for all I know they've already dealt with this. According
>> to the README, we last synced with the 2010c tzcode release, so we're
>> short a few years worth of upstream bug fixes there.
> Agreed- but just to level-set expectations here, don't expect it to be a
> trivial or easily done thing. I looked into a bit when hitting a few
> Coverity complaints (which appear to be innocuous based on both my
> review and that of others who have commented on it) and was quite
> surprised at how far the code has diverged. That code appears to be
> quite actively updated and we may want to try and work out a way to keep
> up better than the occational "whenever someone feels up to it and has
> the time to" updates that we're doing now.
Yeah. IIRC, the existing divergences come mostly from (1) running
pgindent on the zic code, which we could certainly refrain from; and
(2) modifying their code to suppress assorted compiler warnings, for
example they were still not using ANSI C function definition syntax
last I looked. I would be loath to give up (2), so unless upstream
has modernized their code it might be hard to eliminate all the
divergences.
If it weren't for the differing philosophies about compiler warnings,
I'd wish we could create a subdirectory that was *exactly* their
tzcode distribution, in much the same way that the data/ subdirectory
is exactly their tzdata distribution. Then tracking their code updates
would be relatively mindless.
regards, tom lane